Two similar—and equally nonsensical—rumors circulate regularly:
1. Airplanes can (or will soon) fly and land themselves, and
2. Mainframe computers can (or soon will) operate and program themselves.
To disprove the former, view this fascinating video
(a bit less than 15 minutes, worth watching) showing complex—and multi-skilled-person—operation of a current highly advanced plane. To rebut the latter, read mainframe discussion list posts about configuration, programming, operation and debugging issues, which are evolving but hardly abating.
The huge cohort of mainframers considering/nearing/reaching or already in retirement has been compared in crisis scale and inevitability to Y2K. But it differs in many ways. First, it’s not a rigid and universal deadline. Second, staffing changes, skills needed, documentation and resources available differ markedly across companies, industries and regions.
But the common theme is our industry losing the mainframe’s greatest generation (or baby boom)—people who designed, built and supported these systems for nearly 50 years, creating a broad and deep technology ecology.
Accommodating Industry and Generational Changes
Companies react differently to these changes. Some see or use them as justification to abandon mainframes; as always, misleading platform cost comparisons, company politics and not understanding mainframe strengths can lead to unwise decisions. Sometimes amplifying those factors is fear that staffing and skills shortages reduce mainframe long-term viability. Countering this inclination, of course, is the long history of failed efforts to move mission-critical applications and services to other platforms. There’s also inadequate recognition of how mainframe strengths and capabilities have improved over the last decade, let alone since the greatest generation created it.
At the same time, entry paths to mainframe careers have changed. It’s not like the late 1960s when IBM hired hundreds of diverse college grads (math, music, history and more), turned them into beginning programmers, designers, architects through months of full-time classes, then provided on-the-job training during which they learned from masters who’d created first-generation mainframes. These people—and others who learned in other contexts—largely pursued enterprise computing careers, spreading skills and the critical "mainframe mindset" through the industry.
While this isn't a Y2K moment or process, planning and execution is needed to adapt to generational changes. Some sites outsource (to IBM Global Services and other providers); others send work offshore. But many want to remain self-sufficient, fully controlling their IT infrastructure. So managers must consider how to prepare—not for their own retirement, but for ongoing staffing, specifically deciding:
• How candidate successors are identified and recruited
• Whether staff is allowed to train successors
• What resources are available for training successors
Extreme approaches are always shortsighted, whether only hiring experts with decades of experience, or recruiting junior people and expecting them to learn without guidance or training. And it's critical remembering that learning systems—System z technology in general and specific OSs, utilities, networks, diagnosis, etc.—differs from acquiring programming language skills.
Expertise in both areas is most effectively acquired by blending formal training (in-person or distance), self-study, mentoring (preferably with local colleagues but also possible remotely) and constant exposure to best practices (including reading well-written and properly idiomatic code). Especially beware false confidence from knowing programming language details without grasping their interrelationships and operation.
Since languages and everything related to System z evolve, ensure that junior people learn current versions and aren't trained in veterans' outdated habits. Don't let code or configuration reading lead to wholesale copy-and-paste without truly understanding form and function. Monitor progress and don't put people in situations beyond their learning and capabilities; that can lead to immediate problems or—worse— crises taking years to emerge.
It's management's task and challenge ensuring that peers support each other with collaborative systems and program reviews. Even thinly staffed organizations can dedicate occasional lunch hours, brown-bags or pizza breaks to technology sanity checks. And it's easier to quality assure senior staff work when the process is routine for everyone.
Poor communication is a hazard of multi-generational staffs. It's sometimes true that older workers feel that progress stopped with the last system they learned, while younger brethren are sure that the industry (if not the world) began with their first system. So all must be prepared to share their enthusiasms and experience while recognizing value in newer and older technologies.
All of this—reviews, mentoring, communication—only works when corporate culture allows and supports it. If too much outsourcing is done, if senior (usually higher-paid) staff members are encouraged to leave, if there's no respect for the stable System z environment, it's hard to pass on the knowledge, skills and mindset needed to continue.
And yet mainframers don't want to leave their jobs only to see their legacy discarded and rebuilt with duct tape and cheap commodity servers. The mainframe community largely wants to protect its enterprise-computing legacy. So it's important to facilitate generational transition with documentation and staff overlaps.
The good news is that there's powerful-and-growing industry support for this. Decades-old SHARE
provides two immersive conferences every year, uniting mainframe customers, many diverse IBMers and representatives from numerous hardware/software vendors in the goal of preserving, sharing and extending the common knowledge pool. SHARE's zNextGen
project caters to new mainframers with structured and progressive education in overall and detailed technical matters plus the mindset needed to exploit the mainframe's unique industrial-strength capabilities. Smaller scale, regional meetings provide similar benefits to mainframe communities, with less travel and lower cost.
Linking IBM, Industry and Academia
For almost 10 years, Don Resnik has led IBM’s Academic Initiative for System z
—evangelizing with schools, assisting clients and business partners with skills and working with education vendors to create a healthy ecosystem. This unique three-way partnership between IBM, business and academia blends IBM resources, hundreds of schools and educators, and real-world perspective on industry needs.
For much of that time, Marist College in Poughkeepsie, N.Y., has offered extensive onsite and distance-learning mainframe education. Angelo Corridori, Marist's director of large systems education, has presented at SHARE conferences and to IBM Academic Affiliates and supported IBM's yearly Master the Mainframe Contest.
Since Academic Initiative's creation, the workforce has gradually shifted—and continues to shift—from baby boomers to Generation X and Gen Y with a constant trickle of zNextGen students hired from Academic Initiative schools globally.
Larger IBM clients have built internal training programs as well as leveraged external training programs, seeking alternatives similar to Marist College instructor-led online courses, where IBM clients have enrolled close to 1,000 employees for remote, hands-on training.
Corridori considers on-the-job training to be incomplete and inadequate for building a base of knowledge and understanding suitable to build on and expand technically. For example, he initially dissuaded people with several years of z/OS experience from attending an introductory course—but discovered that years of experience are too often just the same year repeated. He notes that formal education provides a solid base, filling in holes.
At the same time, some companies have had great success combining careful hiring and intensive mentoring to build foundational career-long skills and attitudes in "youngsters.”
Resnik sees one of the greatest shifts over the past five years being adoption of "Enterprise Computing" or "Enterprise Systems." He says that this is now academia's preference, used to add new courses and market to students and parents.
A growing number of passionate educators are influencing thousands of students about the need for problem-solvers and creators of technology, as opposed to simply being end users of mobile apps. He predicts that enterprise-computing evangelists will bring forward a new generation of IT workers who understand why the mainframe will celebrate 50 years in 2014.
Additionally, following declining offerings, many vendors are resuming or originating mainframe-related training. Some offer basic/advanced/tailored classes or training for mainframe applications developers.
IBM describes its Innovate 2013
conference (June 2-6 in Orlando, Fla.) as, "IBM's premier networking and educational conference for IT and engineering professionals engaged in all aspects of product development and software delivery." Last year's event had quite a few schools (high school, trade, community college and university) touring students through the show floor featuring mainframe heavy metal; they were reportedly very impressed with large-scale System z boxes.
Encouragingly, IBM offers training in subjects and regions not targeted recently, for example, in Canada presenting z/VM and Linux boot camp with a live virtual classroom option.
As Simple as Possible
So considering multiple various positive trends in maintaining and extending our half-century of mainframe culture, we shouldn't decry simplification of mainframe tasks—or even generational transition—as long as the industry follows advice attributed to Albert Einstein, "Everything should be made as simple as possible, but not simpler." That lets zNextGen mainframers focus on supporting ever-increasing workloads without bogging down in meaningless minutia (no matter how much time we were forced to spend on it).
Gabe Goldberg has developed, worked with and written about technology for decades. He can be contacted at firstname.lastname@example.org.